Genetic defects in PRC2 components other than EZH2 are not common in myeloid malignancies

To the editor:

Several genetic alterations affecting loci encoding epigenetic regulators have been discovered in myeloid malignancies in the last few years. Previously, we and others identified novel mutations in the histone methyltransferase gene EZH2.^1-3 EZH2 encodes the catalytic subunit of the polycomb repressive complex 2 (PRC2), which mediates the methylation of lysine 27 of histone H3 (H3K27). Trimethylation of H3K27 is an epigenetic modification associated with gene silencing. Loss-of-function mutations in EZH2 were found in 6% of patients with myelodysplastic syndromes (MDS), 6% of patients with myeloproliferative neoplasms (MPN) and 12% of patients belonging to the MDS/MPN intermediate group.^1,2 

The PRC2 complex consists of 3 core components, EZH2, SUZ12 and EED, which are required for the complex to execute its function (supplemental Figure 1A, available on the Blood Web site; see the Supplemental Materials link at the top of the online article). A knock-out of any one of these components leads to early embryonic lethality in mice, demonstrating their importance during early development.⁴  Shen et al showed that EZH1, a homolog of EZH2, might substitute for EZH2, because it interacts with the same PRC2 components and likewise needs EED and SUZ12 for its histone methyltransferase activity. Moreover, EZH1 and EZH2 bind to a largely overlapping set of target genes.⁵  Considering these data, we hypothesized that mutations in EED, SUZ12 and EZH1 might be present in patients with myeloid malignancies, leading to a similar phenotype as loss-of-function mutations in EZH2. Furthermore, our recent single nucleotide polymorphism (SNP) array analysis of 102 MDS patients identified one patient with an acquired 5-Mb deletion at 6p22.3 encompassing 15 genes, including JARID2 (supplemental Figure 1B), as well as one patient who displayed uniparental disomy of this region (supplemental Table 1).⁶  As JARID2 can interact with the core components of PRC2 and is involved in the recruitment of PRC2 to specific target genes in embryonic stem cells,⁷  mutations in this gene might perturb PRC2 function as well.

To determine whether mutations in PRC2 components other than EZH2 occur in patients with various myeloid malignancies, we sequenced the entire coding region and splice donor and acceptor sites of SUZ12 (n = 256), EED (n = 326), EZH1 (n = 197) and JARID2 (n = 99; supplemental Table 2). We collected bone marrow and blood from patients with various myeloid malignancies after obtaining informed consent according to the Declaration of Helsinki. This was approved by the medical-ethical committee of the Radboud University Nijmegen Medical Centre. None of the described genes harbored frameshift or nonsense mutations, but we detected 13 novel variants (7 missense, 6 silent) in 20 patients (Table 1). Sequencing of available constitutional (T cell) material showed that 4 of these variants (all missense) were not somatically acquired. For the remaining 9 variants (3 missense, 6 silent), T cells were not available. Among these was a p.E125V variation in EED, which may still be interesting because it is located in a WD40 domain involved in binding to EZH2 and thus could potentially have consequences for PRC2 function. It therefore remains possible that PRC2 components other than EZH2 are mutated, albeit at a very low frequency (< 1%).

In conclusion, our results indicate that, although the PRC2 complex can be impaired in myeloid malignancies because of mutations in EZH2, other important PRC2 proteins are not frequently affected. This suggests that EZH2 is the most vulnerable component of this complex, whose proper function is important for normal hematopoiesis.